The use of different hosts imposes divergent selection pressures on parasitoid populations through fitness trade-offs related to the use of different hosts, thus driving local adaptation of parasitoid populations to their natal host (Kawecki and Ebert 2004). Several interacting factors may counteract the adaptive process, including gene flow, environmental variability, phenotypic plasticity, and the lack of genetic diversity (Kawecki and Ebert 2004; Crispo 2008). The results obtained in this study indicate that natural populations of A. ervi coming from different hosts in the field exhibit important differences in infectivity on their natal host in comparison with non-natal hosts, although this pattern was not observed for all the parasitoid populations nor in all behavioral variables studied. Despite the above, the virulence (a proxy for fitness in parasitoids) expressed by these populations on the tested aphid hosts shows a lack of local host adaptation. Parasitoids obtained from the aphid hosts S. avenae, A. pisum-alfalfa race, and A. pisum-pea race showed a greater infectivity on their natal host in comparison with that shown on non-natal hosts, but only for some of the behavioral variables studied (e.g., time to first sting). For parasitoid populations from S. avenae and A. pisum-alfalfa race, the proportion of time spent stinging was significantly greater when the natal host was offered compared to non-natal hosts, and for the S. avenae-originated parasitoids, the frequency of stinging was significantly greater when the natal host was offered compared to non-natal hosts. Otherwise, significant differences in ‘the time to the first attack’ on hosts were observed for parasitoid populations from S. avenae and A. pisum-pea race, while significant differences for ‘the time to the first sting’ on hosts were observed for three of the four parasitoid populations studied (from both A. pisum races and S. avenae), taking less time to attack and sting when the natal host was offered in comparison with the non-natal hosts. Contrastingly, the virulence assay showed a high plasticity for traits related to fitness. The three different parasitoid populations studied (from both A. pisum races and S. avenae) showed a similar high virulence (parasitism rate, survival, and productivity) on natal and non-natal hosts (APA, APP, and SA), thus providing evidence for the absence of local host adaptation. Additionally, an unusual male-biased sex ratio in the offspring of parasitoid females on both natal and non-natal hosts was observed, as well as an unexpected male-biased sex ratio of parasitoids from the A. pisum-alfalfa race on their natal host when compared to the non-natal host race. A possible explanation for these finding is that aphids from early stages were used for these assays, and inadvertently, some variation in size of the aphids used could have affected, to some degree, the female sex allocation. The theory of sex allocation predicts that parasitoids should lay male eggs in small hosts and female eggs in large hosts (Charnov et al. 1981; Godfray 1994).
The results presented here do not support the ‘sequential radiation’ hypothesis for these introduced populations (Abrahamson and Blair 2008), as parasitoids collected from both host races of A. pisum did not show fitness compromises in the reciprocal transplant experiments (i.e., a higher fitness on their natal host race than on non-natal host races). Consequently, Bilodeau et al. (2013) rejected the sequential radiation hypothesis for introduced populations of A. ervi in North America, based on population genetic and experimental data. The inconsistency between the lack of local host adaptation and the differences in parasitoid infectivity observed is not surprising. A joint evolution of preference and performance is not always observed, due to the need for coupling based on pleiotropy or by linkage disequilibrium between both traits (Futuyma and Moreno 1988; Forister et al. 2007). In addition, unlike the traits related to the parasitoid virulence, behavioral traits related to parasitoids infectivity can be strongly influenced by imprinting effects (i.e., host fidelity due to the preadult-host experience), thus explaining the greater preference to the host from which they emerged. Different studies have demonstrated that host fidelity in A. ervi can be induced even after a single generation, indicating that this mechanism is plastically induced and not under direct selection (Daza-Bustamante et al. 2002; Henry et al. 2008). Although host fidelity has proved to encourage the host adaptation through continual use of the same host species (Forbes et al. 2009), by itself it is not sufficient for the formation and maintenance of host adaptations in parasitoid populations, even under reduced gene flow between populations. Fitness trade-offs on the alternate hosts in these populations is a critical prerequisite for local adaptation (Kawecki and Ebert 2004; Abrahamson and Blair 2008). In this regard, the lack of local host adaptation in our results could be explained by the selective environments where A. ervi is found, where, in rapidly changing environments, generalist plastic phenotypes would be favored. In fact, the evolution of adaptive phenotypic plasticity (i.e., plasticity that increases mean fitness across environments) could be favored over local adaptation in the presence of environmental heterogeneity (Kawecki and Ebert 2004; Crispo 2008; Svanbäck et al. 2009). In this respect, the aphid-parasitoid population dynamics in agroecosystems has been classically described as a metapopulation, characterized by frequent local extinctions and recolonizations (Weisser 2000; Rauch and Weisser 2007). In doing so, frequent extinctions obliterate locally adapted gene pools through increased dispersal, so extinction-colonization dynamics (i.e., metapopulation dynamics) are unfavorable to local adaptation (Kawecki and Ebert 2004). Thus, fluctuations in host abundance (in time and/or space) could favor the maintenance of generalist parasitoid populations with the ability to simultaneously maximize fitness on more than one host (Hufbauer and Roderick 2005). However, other nonmutually exclusive explanations must not be ruled out. For instance, a short time has elapsed since the introduction of A. ervi in Chile, so the potential loss of genetic variation during the introduction process could have delayed the adaptive process. Furthermore, a high gene flow of A. ervi among other aphid hosts could be detrimental to local host adaptation, due to its homogenizing effect of the genetic variation between parasitoid populations through a continuous gene introgression within locally adapted demes (Rasanen and Hendry 2008). In the same sense, a high gene flow could also result during the selection for increased plasticity, or alternatively, plasticity may promote gene flow between different selective regimens (Crispo 2008) by allowing migrants to maintain high fitness (virulence) across alternative hosts and thus facilitate colonization, population growth, and population persistence (Thibert-Plante and Hendry 2010; Chevin and Lande 2011), which would translate into a direct benefit for the control of a target pest.